A Primary Study on Gully Erosion Area Estimation Based on Images

We propose in this paper an approach for image-based gully area estimation. Given an image of a real gully, our goal is to automatically detect gully and measure its area. The main challenge in developing such tool is the lack of camera calibration information for the images. In this research, we have presented a novel approach for gully area estimation. A reference with known area was used in gully area estimation instead of camera calibration, which makes it easier to carry out in the outside gully measurement. The image processing algorithm including segmentation, filtering, edge detection and morphological processing was applied in gully identification. The proposed method was evaluated in the experiment environment. The relative error of the estimation is an average of 2.78%. We also give an example of actual gully detection and gully area estimation using this method, which shows that the method is practical and simple. Thus, it will enable to perform real gully area estimation.